Manganese peroxidase (MnP) has been implicated in lignin degradation and thus has potential applications in pulp and paper bleaching, enzymatic remediation and the textile industry. Transgenic plants are an emerging protein expression platform that offer many advantages over traditional systems, in particular their potential for large-scale industrial enzyme production. Several plant expression vectors were created to evaluate the accumulation of MnP from the wood-rot fungus Phanerochaete chrysosporium in maize seed. We showed that cell wall targeting yielded full-length MnP, whereas cytoplasmic localization resulted in multiple truncated peroxidase polypeptides as detected by immunoblot analysis. In addition, the use of a seed-preferred promoter dramatically increased the expression levels and reduced the negative effects on plant health. Multiple independent transgenic lines were backcrossed with elite inbred corn lines for several generations with the maintenance of high-level expression, indicating genetic stability of the transgene.
Industrial enzymes are widely used in both manufacturing and household products. They often have unique advantages over chemicals with regard to specificity and impact on the environment (Senior et al., 1999; Bhat, 2002; Gupta et al., 2002; van der Maarel et al., 2002). The use of enzymes is growing, but is often blocked by performance issues, the need for large volumes of such proteins and the low cost of chemicals in established processes (Bergquist et al., 2002; Silveira and Jonas, 2002). Industrial enzymes can be isolated from natural sources, but most often they are obtained from recombinant microbial organisms, which facilitate both re-engineering and production of the protein of interest. Although traditional expression systems continue to improve in yield, thus lowering the cost of the enzyme, they can still fall short of cost and/or capacity requirements.
Transgenic plants are an emerging expression platform with enormous potential and unique advantages over traditional sources of protein (Hood and Jilka, 1999; Larrick and Thomas, 2001). Although diverse recombinant proteins can be expressed in many model research plants, such as Arabidopsis, domesticated crop species are more practical because of agronomic characteristics and production advantages. Within the past decade, technologies have been developed to transform virtually all commercial types of plants, including cereal, fruit and vegetable species, tobacco and even trees, with the most common traits amongst such transgenic plants being herbicide and pathogen resistance (Barton et al., 1987; Howe et al., 2002). There is now a growing trend to use such plants for protein production as an end in itself. Maize, in particular, is very attractive because of its high yield, enormous production potential and the stability of its grain, where recombinant protein is often targeted for accumulation. This commodity crop has been the leader amongst plants in providing recombinant proteins to the marketplace. The first commercially produced proteins were avidin and β-glucuronidase (Kusnadi et al., 1998). Of greater commercial significance are bovine trypsin and aprotinin, used for pharmaceutical and industrial applications, both being inherently free of animal pathogens (Zhong et al., 1999; Woodard et al., 2003).
Some proteins are difficult to produce in traditional recombinant systems such as bacteria and fungi. The reasons often include translational and post-translational issues, protein solubility or even enzyme toxicity to the host. Oxido-reductases are an important class of industrial enzymes that can be challenging to express. They have potential uses in detergents, textiles, bio-glues, pulp bleaching and enzymatic remediation (Call and Mücke, 1997; Hüttermann et al., 2001; Lopez et al., 2002). Here, we describe the production of one such enzyme, manganese peroxidase (MnP), in transgenic maize. This enzyme is one of several lignin-degrading enzymes secreted by wood-rot fungi (Kuwahara et al., 1984; Tien and Kirk, 1984; Glenn and Gold, 1985). MnP has been shown to degrade lignin in vitro and to brighten paper pulp (Wariishi et al., 1991; Paice et al., 1993, 1995). The chemical pulping industry currently bleaches 100 million tons of pulp globally per year using harsh oxidizing and bleaching chemicals, leading to an organo-pollutant problem (Call and Mücke, 1997; George Ionides, personal communication). Thus, MnPs and the structurally related lignin peroxidases (LiPs), as well as fungal laccases, have been the subject of intense study with the aim of minimizing the use of bleaching chemicals and polluting effluent (Tien and Kirk, 1984; Archibald et al., 1997; Call and Mücke, 1997). However, to be commercially viable, the enzyme must be produced in very large volumes at a relatively low cost (Paice et al., 1993, 1995; Hüttermann et al., 2001). Despite these challenges, xylanases have been introduced into pulping operations (Senior et al., 1999). Although these enzymes degrade hemicellulosic xylans rather than lignin itself, the net effect of their use reduces the consumption of bleaching chemicals.
MnP is a haem-containing enzyme that requires an oxidant such as hydrogen peroxide to catalyse the oxidation of Mn(II) to Mn(III) (Glenn and Gold, 1985). Once manganic ion is formed, it is removed from the enzyme's active site through chelation by organic acids, forming a soluble oxidant with an unusually high redox potential that is capable of oxidizing phenolic lignin structures (Wariishi et al., 1992). This enzyme has previously been difficult to express in several recombinant systems and haem is often cited as a limiting factor (Whitwam et al., 1995; Stewart et al., 1996; Conesa et al., 2000). Although MnP has been successfully expressed in alfalfa, plant health was negatively affected (Austin et al., 1995), suggesting that other protein expression strategies should be tested.
The goal of this research was to develop transgenic maize for the high-level expression of MnP. To optimize recombinant protein accumulation whilst minimizing potential plant health effects, constructs were created to direct MnP to either the cell wall or cytoplasm, in combination with either constitutive or seed-preferred promoters. Our results showed that significant levels of the active enzyme accumulated in the seed if a seed-preferred promoter was used and secretion was directed to the cell wall. In addition, we demonstrated the genetic stability and continued high-level expression of recombinant MnP in transgenic lines over the course of three generations in the field.
The coding region of the MnP1 gene from Phanerochaete chrysosporium was obtained from pTAAMnP1, a plasmid engineered for MnP expression in Aspergillus (Stewart et al., 1996). Plant transformation vectors were designed for cytoplasmic (MPA) and cell wall (MPB) targeting of MnP under constitutive control (Streatfield et al., 2004). MPB contained the barley α-amylase signal sequence in place of the native signal sequence (Rogers, 1985). To mitigate any potential negative effects on plant health, we also tested a seed-preferred promoter (Belanger and Kriz, 1991), included in construct MPD (Figure 1).
Herbicide-resistant transgenic calli (T0 generation) were obtained from each plant vector (Table 1). For each independent transgenic event (event), 10 sibling plantlets (lines) were transferred to soil, grown in a glasshouse and outcrossed with pollen from elite inbred lines. As the expression of MnP has the potential to affect plant health, we noted the number of non-viable ears (ears producing no seed after pollination) and seed yield. Plants transformed with MPB exhibited a relatively high rate of non-viable ear production (Table 1) compared with MPD transformants. A few MPB events produced no seed from any line (data not shown). In addition, plants transformed with MPB, in particular, displayed early leaf senescence and high rates of male sterility (data not shown). Nevertheless, if ears were produced, the average amount of seed was similar to that obtained from plants transformed with MPA, whereas, in comparison, MPD plants produced nearly twice the amount of seed.
Table 1. Transformation results
Construct (promoter, target)
Number of events
Average no. of seed per ear
Fraction of events positive for MnP
At least 10 sibling lines per event (independent transgenic event) were transferred to the glasshouse. The fraction of non-viable ears (NVE) was based on the number of plants (np) surviving until pollination with elite maize. The average number of seed per ear was based on the number of ears (ne) harvested. Analysis of manganese peroxidase (MnP) activity in T1 seed failed to detect any significant activity in any MPA events. Activity could be seen in all events from MPB and MPD; however, note that all plants from three MPB events failed to survive.
MPA (constitutive, cytoplasm)
11.2 (np = 152)
59 (ne = 135)
MPB (constitutive, cell wall)
27.6 (np = 116)
55 (ne = 84)
MPD (seed-preferred, cell wall)
4.4 (np = 137)
105 (ne = 131)
MnP expression in T1 seed
To test for the expression of MnP in transgenic maize, protein extracts were prepared from the individual seed from T1 ears, representing several events, and analysed by immunoblot and activity assay. Using an MnP-specific antibody, a band of approximately 46 kDa was detected on analysis of lines from both MPB and MPD (Figure 2). The band was similar in size to authentic MnP from P. chrysosporium, which was included as a positive control. The absence of bands in some lanes is consistent with the outcross of T0 plants and a single locus insertion. These samples were also assayed for MnP activity, which is shown at the bottom of the blots (Figure 2a,b). MnP activity was seen only in samples in which a putative MnP band was detected. Little or no activity could be detected in extracts from MPA events, and the putative MnP band observed for both MPB and MPD was absent on the MPA blot (Figure 2c). Screening of T1 seed from 70 ears representing 15 independent MPA events failed to identify a single ear with significant activity (Table 1). Although full-length MnP could not be detected on immunoblots, two smaller bands (∼30 kDa and 25 kDa) were evident and are indicated by the lower arrows (Figure 2c).
Peroxidases are present in many plant tissues and could potentially interfere with high-throughput screening of transgenic plants and measurement of the recombinant enzyme. Therefore, we examined the level of background activity present in several inbred lines that were components of the genetic background of these transgenic plants, including non-transgenic HiII/elite ears. Analysis of protein samples from the various non-transgenic maize plants showed little difference in activity between samples with and without manganese in the assay (Figure 3 and data not shown). In contrast, high levels of manganese-dependent activity were seen in samples derived from plants transformed with the seed-preferred construct, MPD (Figure 3). Results are shown for T1 event 01, including samples from two different lines and a null seed from line 0108 (see Figure 2b for immunoblot). Similar results were seen for all MPB events tested (data not shown).
Identification of transgenic maize lines expressing high levels of MnP
To compare the relative effectiveness of each construct and to determine which events to select for further assessment, the MnP content in T1 seed was measured. The amount of MnP in protein extracts was calculated as a percentage of total soluble protein (TSP) after measurement of activity and comparison with known amounts of recombinant enzyme (Whitwam et al., 1995). MnP levels are shown for the individual seed with the highest MnP content out of five seeds analysed per ear (Figure 4). In cases in which the amount of MnP in samples exceeded the reference level, the second highest quantitative value is also shown. Just as there was a variation in MnP levels between seeds from the same ear (Figures 2 and 3), variation was seen between lines within the same event. For example, a greater than 10-fold difference in MnP levels was evident between lines within the same event. Our goal was not to determine the cause of variation between samples and lines within an event, but to select events with high MnP content for further analysis. Every MPD event produced lines exceeding at least 6% TSP as MnP. The majority of MPD events generated lines with MnP levels of at least 11% TSP, with the top lines being over 14% TSP. In contrast, the best events generated with the MPB construct were in the range of 2%−3% TSP. Therefore, when comparing the majority of T1 lines, the data suggested that higher levels of MnP could be expressed in seed when using a seed-preferred promoter rather than one that was constitutive.
Phenotypic effects of MnP
Phenotypic data from T0 plants suggested that constitutive expression of MnP activity, especially if targeted to the cell wall, was detrimental to plant health. To determine whether these effects were a result of MnP or environmental conditions, T1 plants from MPA, MPB and MPD were grown simultaneously in the glasshouse. Lines with the highest amount of MnP and which had been outcrossed to the same inbred line were selected from three events per construct. Ten seeds per line were planted. After the selection of plants with herbicide (leaf painting with bialaphos) to identify those that had inherited the transgene, two individuals per line were selected for further analysis of growth and development. Prior to tassel development, no obvious phenotypic differences between plants were observed. However, shortly thereafter, small, oval-shaped, rusty-coloured lesions began to develop on the surface of the most mature leaves on MPB plants (Figure 5). This was not seen on the MPA or MPD plants. Figure 5 shows several examples of leaf and stem symptoms. Leaves were thoroughly examined for evidence of fungal growth using a dissecting microscope, but none was found. Of particular note was the eventual development of lesions on the stalk. The severity of the leaf phenotype resulted in premature senescence of the leaf tip, followed by the outer edge of the blade, developing towards the vein. Foliar and stem symptoms increased with leaf age. Moreover, although ears were obtained from all MPA and MPD plants (12 ears), only three of the six MPB plants produced ears. Furthermore, the symptoms were more severe for one of the two individuals in each event, and, in each instance, only the healthier plants produced ears.
Field performance of MnP-expressing transgenic maize
To assess agronomic performance, inheritance and expression of MnP in transgenic corn, T1 lines representing at least two MPB and MPD events were selected for planting in the field. Seedlings were leaf painted with bialaphos to screen for progeny that had inherited the transgene. In the field nursery, just as in the glasshouse, plants transformed with the MPB construct exhibited phenotypic abnormalities within leaves and many of these plants did not produce any seed. This was observed for each event and did not seem to correlate with the expression level of MnP in the parental seed (data not shown). Because of the higher expression levels and superior health of MPD plants, further evaluation was not continued for MPB. Effort focused on lines from the top two events (MPD01 and MPD08) and two alternative events (MPD03 and MPD12). As in the glasshouse, MPD growth and development were excellent. We did not observe any phenotypic anomalies as seen with MPB transformants in the field. Various pollinations included backcrosses to transfer the MnP gene into different elite inbred lines, self-pollinations and outcrosses to high-oil lines to test for increased MnP content. Ears were harvested and MnP was measured in protein samples derived from pools of 50 seeds per plant. Figure 6 summarizes the MnP content measured in T2 ears obtained from the various pollinations. Data are shown for individual ears and, if the average is given, one or two of the highest values are also included to provide insight into the potential upper limit. With the exception of MPD0103, ears from self-pollinations contained the most MnP. This is to be expected as more copies of the transgene should be present in self-pollinated as opposed to backcrossed ears. The average level of MnP resulting from self-pollinations for lines 0108 and 0802 was approximately twice that seen when the same lines were backcrossed. The highest MnP content within events (other than self-pollinations) resulted from crosses with high-oil rather than elite lines, but again MPD0103 was the exception.
Transgenic plants were advanced to the field nursery for two more generations to study both the inheritance and expression of MnP. Figure 7 shows the performance of MPD01 lines backcrossed to either elite or high-oil lines. To estimate the production potential from grain, the amount of enzyme was calculated in terms of seed weight. Just as in the T2 generation, use of high-oil germplasm appeared to be very effective at increasing the MnP content in T3 and T4 seed relative to backcrosses with elite lines. Within each season, the amount of MnP per dry weight in ears resulting from crosses with high-oil corn was approximately twofold greater than that in ears resulting from crosses with elite inbreds. Furthermore, we did not observe any negative effect on plant health.
The commercial potential of fungal oxido-reductases has previously been demonstrated by their effectiveness in bench-scale application trials, such as in paper pulp bleaching. However, manganese-dependent peroxidases, in particular, have not been developed for industrial use, in part because they do not express well in recombinant systems. Fungal fermentations can be highly productive, yielding gram quantities of protein per litre. However, only 5–100 mg per litre could be obtained using Aspergillus, which required supplementation of media with large amounts of haem (Stewart et al., 1996; Conesa et al., 2000). Given our previous success with using transgenic maize to produce industrial enzymes in seed, our goal was to demonstrate the feasibility of the use of this system for the production of MnP.
Previous attempts to express MnP in plants have shown a direct correlation between expression level and stunted plant growth and development (Austin et al., 1995). Harmful effects may be the consequence of constitutive targeting of the enzyme to the cell wall where there are many potential substrates, including lignin. Therefore, in addition to testing a construct (MPB) similar to those used in previous studies, we chose to target the enzyme to the cytoplasm to remove it from potentially reactive apoplastic substrates. To avoid problems linked to constitutive expression, a third vector (MPD) was created to limit cell wall expression preferentially to the seed.
Unfortunately, although we recovered healthy plants transformed with MPA, extensive enzymatic screening failed to recover any event with significant enzymatic activity. Instead, using Western blot analysis, two bands of approximately 25–30 kDa were seen, possibly representing degradation products. In contrast, full-length and enzymatically active MnP accumulated in seed from plants transformed with MPB. However, in the T0 generation, there were some signs of poor plant recovery, but it was difficult to distinguish between effects linked to MnP as opposed to those caused by tissue culture (transformation) or environmental factors. For example, seed yield was quite similar between MPA and MPB plants, but the number of non-viable ears was greater amongst MPB events (Table 1). In contrast, the use of a seed-preferred promoter (MPD) seemed to eliminate plant health problems in T0 plants, even though MnP was targeted to the cell wall and accumulated relatively larger amounts of enzyme than MPB events.
To help clarify whether plant health problems were in fact construct specific, MPA, MPB and MPD plants were grown under the same conditions at the same time. We observed a striking lesion phenotype on leaves and stalks with the onset of flowering. This was seen only amongst plants transformed with MPB and was never observed amongst MPD plants, even in the highest expressing lines. Furthermore, those plants with the most severe symptoms failed to survive and produce seed (one of two transgenic plants per event). Interestingly, even though the MnP content was high in MPB seed, we did not observe any effect on germination or plant development before flowering. Thus, phenotypic effects were not linked to high levels of accumulation in the seed. Figure 5 shows that symptoms developed in older leaves first. Constitutive expression with targeting to the cell wall was also problematic in alfalfa (Austin et al., 1995). Plants were stunted with delayed flowering and leaves turned yellow, followed by senescence, without forming lesions. These authors also showed that plants with MnP levels greater than 0.3% TSP in the leaf rapidly senesced and died. Finally, using a similar construct, it was possible to generate healthy tobacco plants expressing the MnP gene from Coriolus versicolor (Iimura et al., 2002). The differences between these studies may be explained by the use of the cauliflower mosaic virus (CaMV) 35S promoter and/or different MnP genes, or perhaps by the use of tobacco. It should be noted that only relative amounts of active enzyme were reported, making it difficult to assess whether the expression in leaf was high enough to cause a phenotype.
As the goal of this study was to develop maize for the high-level expression of MnP, T1 plants were grown under field conditions to further evaluate the inheritance and expression of the transgene as well as agronomic performance. MnP levels above 10% TSP were seen in self-pollinated ears from events 01 and 08, whereas ears from backcrosses generally contained half as much MnP, ranging from 3% to 6% TSP, possibly due to zygosity. Previous studies on the fungal production of MnP have shown that the expression level is tightly connected to the availability of both iron and haem, which make up the enzyme's active site prosthetic group (Stewart et al., 1996; Conesa et al., 2000). The results presented here suggest that these molecules are not limiting in transgenic corn at current expression levels, and that it may be possible to further increase the enzyme content through the selection of high-expressing lines or simply by creating homozygous plants. Finally, the MnP content was consistently higher in ears resulting from backcrosses with high-oil lines. In the T4 generation, seed contained nearly 500 mg MnP/kg, whereas the highest level resulting from backcrosses with elite lines was less than 200 mg MnP/kg. The benefits of using high-oil lines have been shown previously for laccase, a highly reactive fungal oxido-reductase. Not only were the seed levels of laccase at least twice as great as in elite lines, but also poor germination frequency was partly eliminated (Hood et al., 2003).
Previous attempts to produce MnP in plants for industrial purposes have focused on expression in leaf and root. Our goal was to produce enzyme in the seed, which is possible when using a constitutive promoter, but clearly difficult to achieve in the field unless a seed-preferred promoter is employed. To examine the feasibility of producing enzyme for commercial purposes, we reviewed the application trials that derived enzyme from traditional expression platforms. Whatever the source of recombinant protein, its efficacy is usually evaluated first at the bench scale. For example, in pulp bleaching experiments, milligram quantities of enzyme were found to be sufficient for analysis. Such studies were performed using crude fungal extracts containing MnP at approximately 5 µg/mL (Paice et al., 1993, 1995; Archibald et al., 1997). To our knowledge, the highest level of MnP reported for fungal fermentations was 100 µg/mL using Aspergillus niger, a significant improvement over previous reports, but less than the gram quantities often achieved with other industrial enzymes (Conesa et al., 2000). Extracts from T3 and T4 maize in this study contained MnP at a concentration of 25 µg/mL and higher, clearly within the range necessary for analysis. Thus, maize may be an alternative to traditional fungal fermentation systems in which expression is inefficient, especially in the situation in which many kilograms of enzyme are needed, such as in pilot trials. In addition, the maize expression system described here may also facilitate the production of other difficult to express oxido-reductases, such as the LiPs. In this regard, the recently published genome from P. chrysosporium reveals a rich source of multiple MnP and LiP genes (Martinez et al., 2004).
The data presented here demonstrate that high levels of MnP can be produced in the field without affecting plant health. This will greatly facilitate grain production; however, future experiments must be aimed at showing that corn-derived MnPs are effective in application trials.
Plant expression vectors
Plasmid pTAAMnP1, containing an MnP coding sequence from P. chrysosporium, has been described previously (coding sequence 100% identical to GenBank accession no. L29039; Stewart et al. 1996). Plant expression vectors utilized only the MnP coding sequence minus the native signal sequence, which was replaced by the barley α-amylase signal sequence (BAASS; Rogers, 1985; GenBank accession no. K02637). The sequences of BAASS and the stop codon were changed to reflect optimal codon usage in corn. The Signal P software tool (http://www.cbs.dtu.dk/services/SignalP/) predicted that BAASS would be cleaved to produce the native polypeptide.
The coding sequence for MnP was amplified by polymerase chain reaction (PCR) using pTAAMnP1 as template to either remove the native signal sequence or replace it with the sequence for BAASS. Both PCRs included a reverse primer encoding an HpaI site for cloning, followed by the stop codon and 3′ coding sequence. Both forward primers encoded an NcoI as part of the start codon, followed by sequencing to either remove or replace the native signal sequence with that of BAASS. The resulting NcoI-HpaI fragments, encoding either MnP or BAASS:MnP, were ligated with the vector 2774 digested with BbsI-HpaI, which contains a maize polyubiquitin-like constitutive promoter (Hood et al., 2003; Streatfield et al., 2004) and the Pin II terminator sequences, resulting in plasmids K2686 and K2704, respectively. The HindIII-NcoI promoter fragments from both K2686 and K2704 were removed and replaced with the HindIII-NcoI fragment from 7583, which contained the constitutive promoter, resulting in K2792 and K2781, respectively. HindIII-NotI fragments from both K2792 and K2781 were then ligated into the HindIII-NotI binary vector fragment from 8916, resulting in vectors MPA and MPB, respectively. The NcoI-NotI fragment from K2781 (see above description) containing BAASS:MnP and the HindIII-NcoI fragment from KB381 containing a globulin-1 seed-preferred promoter (Belanger and Kriz, 1991) were ligated into the HindIII-NotI vector fragment from 8916, resulting in vector MPD. A maize codon-optimized phosphinothricin N-acetyltransferase gene (moPat) from Streptomyces viridochromogenes was present on plant vectors to confer glufosinate resistance (White et al., 1992). Constructs were sequenced to confirm that no errors had been introduced during the cloning procedures.
Plant vectors were transformed into maize using an Agrobacterium tumefaciens-mediated transformation procedure described previously (Hood et al., 2003). Multiple lines representing each independently transformed event were sent to the glasshouse for pollinations by elite inbreds and to recover T1 seed. Southern blots were used to confirm integration of the MnP coding sequence in lead lines.
Protein extracts from transgenic maize samples were prepared by either pulverizing single seed or by grinding 50 seeds per ear in a coffee grinder, followed by extraction with 50 mm sodium tartrate, pH 4.5. The protein concentration of the extracts was determined by the method of Bradford, using bovine serum albumin (BSA) as a standard (Bradford, 1976). MnP activity in the extracts was measured by monitoring the oxidation of 2,6-dimethoxyphenol at 469 nm (Wariishi et al., 1992); 0.2–1 µg of seed protein extract was assayed at 28 °C for 5 min in 50 mm tartrate, pH 4.5, containing 0.5 mm manganese sulphate, 1 mm 2,6-dimethoxyphenol and 0.05 mm hydrogen peroxide. MnP standard was obtained from Dr Ming Tien at Penn State University, Pennsylvania, USA. Levels of MnP in the extracts were reported as the percentage of TSP, or on a dry weight basis (whole flour). To identify MnP by immunoblot analysis, proteins were separated by sodium dodecylsulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto immobilon-P. Blots were probed using antibodies raised against MnP isozymes and followed by chemiluminescence detection.